Emotional discomfort among school staff and student teachers when they had to 'tell' or 'not tell' about vague suspicions of domestic violence.

BACKGROUND: Many employees in schools and kindergartens fail to report vague suspicions (e.g., inadequate clothing, signs of anxiety or lack of sleep, being verbally violent to peers) that children are victims of domestic violence because they are unsure whether their suspicions are well founded.
OBJECTIVE: We investigated the degree of emotional discomfort among school staff and student teachers when they had to 'tell' or 'not tell' about vague suspicions of domestic violence. We assumed that they would experience more emotional discomfort when they did not report such vague suspicions. PARTICIPANTS AND
SETTING: Seventy-one teachers and student teachers (20 men and 51 women; average age: 25, age range: 18-62) were recruited from three primary and lower secondary schools.
METHODS: We used an experimental design in this quantitative vignette study, and the particiants were asked to respond to 54 statements that were related to the need for acceptance by others, seriousness, appraisals (e.g. social self-image and self-image) and self-critical feelings (e.g. rejection, shame, and inferiority).
RESULTS: The results of the experimental study show that participants who did not tell about their vague suspicions of violence reported this as being more serious (p < .05), had a more negative self-appraisal (p < .05) and reported a higher degree of negative feelings (p < .05). The study sheds light on the seriousness of 'not telling' about vague suspicions, and shows that 'not telling' is reported as being more emotionally uncomfortable than 'telling'.

Introduction

Violence is a social problem that is detrimental to both the economy and people's health, and domestic violence accounts for one in six violent crimes (Statistics Norway, 2017). A total of 3,450 instances of domestic abuse were registered in 2016 (Statistics Norway, 2017). Violence can have serious short-term and long-term consequences for many children (Gamst, 2017). For children who have experienced domestic violence, the consequences can affect their ability to adapt, their ability to form attachments, their emotional regulation skills and behaviour (Kirkengen and Næss, 2015). Disruptions to one or more of these factors can lead to maladjustment during a child's development and give rise to cognitive, emotional, behavioural and social problems (Øverlien and Moen, 2016; Mossige and Stefansen, 2016).

Children and young people who witness violence, for example in the home, are often also directly subjected to domestic violence (Nordhaug, 2018). According to Sjøvold and Furuholmen (2015), many children who are subjected to domestic violence have poorer mental health, struggle with social relations, have attachment problems and could potentially develop Post Traumatic Stress Disorder (PTSD).

Violence is defined as: ‘Any action directed against another person that, because this action is capable of injuring, inflicting pain, frightening or violating, causes that person to do something against his or her will or to cease doing something he or she wants to do’ (Isdal, 2018, p. 7). Isdal defines several different types of violence, which he has divided into five sub-categories: physical violence, psychological violence, sexual violence, material violence and latent violence (Isdal, 2018). In our study, we have focused in particular on vague suspicions of violence, which, according to Isdal (2018), can be challenging since violence can be expressed in many different ways and have many different symptoms. This may be, for example inadequate clothing, signs of anxiety or lack of sleep, being verbally violent to peers or personnel, reactions of fear towards sudden movements.

To succeed in preventing children being subjected to domestic violence and the consequences it brings, it is important to be able to detect the violence. Schools and kindergartens (i.e., the educational system) are important arenas in this context. School and kindergarten staff follow children for long periods of time and have an opportunity to observe changes in their behaviour and development (Danielsen et al., 2016). Many employees in the educational system feel that they do not have sufficient expertise to deal with an issue of this kind, and therefore do not ask children about it even though they have vague suspicions of domestic violence (Øverlien and Moen, 2016). The consequence is that many children live with domestic violence without being detected by the system. Øverlien and Moen (2016) write that eight out of ten primary and lower secondary teachers feel that they did not learn enough about domestic violence during their teacher training. Children often have great trust in school and kindergarten staff. Children who are subjected to violence often state that they are disappointed that their teacher has not talked to them or asked them if they are being subjected to violence (Øverlien et al., 2015).

Violence can be difficult to detect and, according to Hage (2014), teachers who report suspected violence often experience that their suspicions are not followed up. They feel that the process is time-consuming and of little or no use (Hage, 2014). According to Kenny (2001), some teachers did not report suspicions of violence because they were worried that they did not have enough information to write a detailed report. The fear of making mistakes thus had a stronger pull than the statutory duty to protect children at risk (Kenny, 2001). Uncertainty about their suspicions is another reason why many teachers do not report suspicions of violence against children (Talsma et al., 2015). Webster, O'Toole, O'Toole and Lucal (2005) show that knowledge about violence and abuse, experience of submitting notifications of concern to the child welfare service, and knowing that it is right to report such things, are important factors that influence school staff's decisions on whether to submit a notification of concern (Webster et al., 2005).

There can also be challenges relating to concern about making a mistake, and thereby running the risk of giving offence or triggering a negative reaction from the family of the child in question (O'Sullivan, 2009). Moreover, this can be related to concern on the part of the person reporting suspicion about his or her social self-image and that the family in question or others will dislike him or her.

Communicating about such serious topics can also affect the self-image of the person reporting a vague suspicion, which can lead to self-critical feelings such as shame, inferiority and rejection (Buckman, 1984). Tesser and Rosen (1972) also show that people are generally reluctant to communicate negative information as opposed to positive information. Taken together, this can give rise to an emotional discomfort that can affect those who plan to report a vague suspicion that a child is being subjected to domestic violence. The Norwegian laws is clear about what must be done where violence is suspected, and the duty to report to the child welfare system. When vague suspicions of domestic violence are not reported, the result may be that the child in question is not extricated from the difficult situation and has to go on living with violence.

In this article, we endeavour, through an experimental study, to examine emotional discomfort in school staff when they have to report, as opposed to not reporting, vague suspicions that children are being subjected to domestic violence. In some previous studies, Torp Løkkeberg (2016) found that emotional discomfort with feelings such as shame, inferiority and rejection play an important role in how unpleasant information is communicated. The Study on which this article is based is a follow-up to this work. It is very important to uncover how self-critical emotions affects school-employees experience vague suspicions of domestic violence in order to raise awareness of how this can affect communication.

Violence and the grey areas

Violence in the grey areas refers to diffuse and less obvious symptoms of violence (Isdal, 2018). Diffuse symptoms can include either a lower or higher level of activity than usual, that pupils withdraw from friends and take no interest in things they have previously been interested in (Baker et al., 2002). Symptoms can be complex and unclear, which makes them difficult for school staff to detect (Dybsland, 2007). According to Danielsen et al. (2016), not enough is known about this topic, both in the teacher training context and in work with pupils in general. Several municipalities in Norway, have established interdisciplinary consultation teams working with children, comprising representatives of the child welfare service, the police, the local health centres and the child and adolescent psychiatric service, to ensure that several agencies cooperate to detect violence. Interdisciplinary cooperation is therefore very important if children's best interests are to be safeguarded.

The requirement that schools must report violence

Domestic violence is a criminal offence, and the law emphasises that employees must report vague suspicions that children are being subjected to violence (The Penal Code, 2005; The Child Welfare Act, 1993; The Convention on the Rights of the Child, 2003). Public authorities must inform the municipal child welfare service when there are suspicions that children are being mistreated at home or when serious neglect is suspected (The Child Welfare Act, 1993). School and kindergarten staff have a duty to report the matter to the child welfare service if they suspect that children and adolescents are being subjected to domestic violence. The duty to report applies irrespective of what their profession is (The Child Welfare Act, 1993). Several national awareness campaign has been held to enlighten public authorities their responsibility to report also their vague suspicions of violence. According to Norwegian laws it is consider as a crime not to report.

The importance of social bonds when communicating unpleasant information

Communicating unpleasant information, for example by reporting vague suspicions of violence, can potentially have a negative effect on social bonds. A social bond ‘involves mental and emotional attunement between people’, according to Scheff (1994, p. 201), so that everyone feels appreciated and respected. Communicating negative information can threaten people's fundamental need for acceptance and belonging, which can lead to people withholding the negative information in order to not break a social bond (Scheff, 1994; Maslow, 1970). The experience of damaging a social bond with another person whom one regards as important can be emotionally stressful for an individual. For instance, the social bonds with the child and the family can be damaged if an employee reports vague suspicions of violence in the home. If a school staff member chooses to withhold negative information in order to protect the social bond, he or she nonetheless risks the social bond being damaged, for example if someone who later receives this information learns that the staff member has withheld it (Scheff, 1994). On the other hand, if the staff member chooses to share the unpleasant information, he or she may receive a negative response from the recipient, which, in turn, can lead to a weakening of the social bond (O'Sullivan, 2009).

Based on the theory of the importance of social bonds, we believe that school and kindergarten employees will wish to protect their social bonds with colleagues, children and the children's families. It is conceivable that reporting vague suspicions of violence could be a threat to employees' social bonds with such other parties. An employee may be concerned about being frowned upon by colleagues, the child and the child's family, particularly if the suspicion proves to be unfounded. On this basis, it would seem that some employees will choose to report vague suspicions, while others will perhaps choose not to (Talsma et al., 2015). Still, there are other factors that influence the underreporting on children exposed to domestic violence (e.g., law enforcement failure, lack of protocol, lack of operational definitions of "violence in grey areas", lack of training).

Appraisals

How people appraise situations predicts different feelings and affects how people are motivated to cope with them. This helps to explain how feelings arise, and how situations are interpreted based on previous emotional experiences (Bippus and Young, 2012). Subjective appraisals are based on feelings arising because of subjective advance assessments of the situation. People assess how a situation will unfold, and feelings arise as a result of these assessments, not as a result of the situation itself (Bippus and Young, 2012). What feelings arise is highly individual and depends on the person's cognitive processing of experience (Ellsworth and Scherer, 2003). Reactions to reporting vague suspicions of violence could be one example of this. Some employees will perhaps feel relieved about having told someone, and feel that they may have contributed to helping a child in a difficult situation. Other employees may have feelings of guilt and shame about having reported their suspicions, because it can have negative consequences for the employee afterwards if the suspicions prove to be unfounded.

When a school employee has vague suspicions that a child is being subjected to domestic violence, he or she faces a dilemma (i.e. to report their suspicion immediately or to put off reporting it). The choice can depend on several factors. The first factor concerns whether the individual risks something in the situation. For example, there may be a conflict between several of the individual's values. An employee can have vague suspicions that a child is being subjected to domestic violence. The individual in question may set great store by values such as honesty and trust, but at the same time have values such as not hurting or upsetting other people. These values can come into conflict with each other when the individual has to choose whether to report or not report such suspicions. The second factor concerns whether the threat to these values affects the individual's wellbeing (Lazarus, 1991; Thiel et al., 2011). For instance, the employee can experience stress if reporting a vague suspicion hurts others or causes them emotional discomfort (Lazarus, 1991). According to Gausel and Leach (2011), the third factor concerns how an individual perceives the situation in relation to his/her own self-image, for example: ‘I, who do not report vague suspicions of violence, have a specific defect’. This also concerns social self-image, for example: ‘I may be disliked by others because I do not report vague suspicions of violence’. For an employee, the situation can be perceived as a threat to his/her personal values, which leads to concern on the employee's part about his/her self-image and social self-image (Gausel and Leach, 2011).

Feelings

Self-critical feelings are triggered in situations involving moral dilemmas and when people are uncertain about what is the right or wrong thing to do (Stiegler, 2013). Feelings such as shame, inferiority and rejection are especially important (Gausel and Leach, 2011). This can apply to situations where employees are considering whether or not to report vague suspicions of violence. Shame is regarded as a very complex emotion, and it can be extremely painful, but also healthy and necessary in the sense that it protects and regulates relations (Farstad, 2016). In some cases, shame can also motivate people to repair relations, or to acknowledge that they have hurt the other party in a relationship, and thereby contribute to the individual who is experiencing shame repairing the social bond (Gausel and Leach, 2011).

Felt inferiority is more about the person in question perceiving his/her whole self as a failure. It is often described as a key feeling among people with mental illnesses, for example depression (Gausel et al., 2012).

Rejection is also a self-critical emotion that is very relevant to understanding emotional discomfort. Gausel and Leach (2011) write that rejection is a feeling of both physical and psychological exclusion or isolation from people who devalue you (MacDonald and Leary, 2005; Scheff, 2000; Gausel and Leach, 2011). People who feel rejected can therefore withdraw or become defensive in order to avoid further or future rejection (Gausel and Leach, 2011). Feeling rejected can strongly affect the social bond because the person in question is motivated to withdraw (Gausel and Leach, 2011).

The objective of this article is to explain the emotional discomfort experienced by those who work in, or train for a job in, the educational system when they have to ‘tell’ or ‘not tell’ about vague suspicions of violence against pupils. Our point of departure was to examine whether there are differences between the participants in relation to the two conditions. This is in order to explain what emotional effect the different conditions (i.e., tell, not tell) can have on a person who has such a vague suspicion.

Methods

Choice of design

We have used an experimental design in this quantitative vignette study. The advantage of a vignette study is that the social stimuli, which in this case are the vignette and the different conditions, are standardised, and that it is more realistic to respond to the different statements in the questionnaire than in a normal questionnaire (Brink and Wood, 1998). Another advantage of vignette studies is that the conditions in the independent variable are kept constant across a group of participants, and they also ensure greater control of stimuli (Brink and Wood, 1998).

Hypotheses

We expect there to be differences in how the participants perceive themselves, depending on which condition they are assigned to. Based on previous studies of communicating unpleasant information, we assume that not telling about vague suspicions will cause more emotional discomfort then telling will (Torp Løkkeberg, 2016). Moreover, we assume that the participants will regard not reporting vague suspicions of violence as more serious than reporting them.

Participants and setting

Seventy-one teachers and student teachers (20 men and 51 women; average age: 25, age range: 18–62) were recruited from three primary and lower secondary schools. In total, there were 35 teachers and 36 student teachers. They were also contacted personally in canteens and libraries at university colleges and universities in Eastern Norway. We sent information letters to three primary and lower secondary schools in the eastern part of Norway. They were recruited as they were part of our university's partner schools. All the participants were randomly assigned into the two groups, and there were no demographics differences between the groups.

The respondents have participated voluntarily, and they have not received compensation of any kind. It was important to us that the participants responded to the questionnaire survey based on their own motivation and not because of external incentives. Only three respondents chose to withdraw from the study before they had completed the questionnaire. Four questionnaires were excluded from the analysis because the respondents had only completed the first page. Personal data were not collected, except for demographic data about age, sex and education. The participants were ensured anonymity and confidentiality. Our Study was conducted in accordance with The Research Council of Norway's privacy considerations and ethical principles.

Structure of the questionnaire

In the first section of the questionnaire (page 1), we asked the participants to fill in demographic data such as age, sex and ongoing education. The participants were asked to envisage the following scenario: Imagine that you have a vague suspicion that one of the children in your class is being subjected to violence (e.g. physical, psychological, sexual). The participants who had been given the questionnaire containing condition 1 read the following: ‘you decide to tell about your vague suspicion’ (N = 38). The participants who had been given the questionnaire containing condition 2 read the following: ‘You decide not to tell anyone (the school nurse, a colleague, a child welfare service employee) about your vague suspicion’ (N = 33). Moreover, they were asked to repeat what they had been asked to envisage, thereby confirming that they had understood the vignette text and to give an example of the situation they imagined. This functioned as a manipulation check (Brink and Wood, 1998). Immediately afterwards, they were asked to respond to 54 statements that were related to the need for acceptance by others, seriousness, appraisals (e.g., social self-image and self-image) and self-critical feelings (e.g., rejection, shame, and inferiority). On average, it took 20 min to complete the questionnaire. The participants responded to the statements using an intensity scale ranging from 1 (not at all) to 7 (strongly agree). There were no demographic differences between the two groups, except from group 1 had a higher number of men n = 15, compared to group 2 n = 5, and group 2 had an age range 18–54, compared to group 1 age range 20–62.

The instrument

The statements used in the questionnaire were inspired by an already validated questionnaire created by Gausel and Leach (2011); Gausel et al. (2012): Torp Løkkeberg (2016). Previous studies has shown alpha values above .70 on the variables used in this study. In this Study, all our variables have alpha values above .70 (Tavakol and Dennick, 2011).

Acceptance

The need for acceptance (α = .94) was measured by eight statements and divided into ‘acceptance by colleagues’ (e.g., I want my colleagues to like me) and ‘acceptance by the family’ (e.g., I want the child's family to like me).

Concern about ‘social image’ (α = .90) in relation to colleagues was measured using two statements (e.g., Colleagues may dislike me because I told/did not tell about my vague suspicion). Concern about ‘social image’ (α = .94) in relation to others' was measured using two statements (e.g., Others may dislike me because I told/did not tell about my vague suspicion). Concern about ‘self-image’ (α = .73) was measured using two statements (e.g., That I (did not tell) told about my vague suspicion revealed a moral failing in myself).

Seriousness

How the participants perceived the seriousness (α = .98) of the situation they were asked to envisage was measured by four statements (e.g., It was wrong (not) to tell about my vague suspicion).

To measure ‘Shame’ we used three statements (e.g., I feel ashamed when I think about how I told (did not tell) about my vague suspicion). We measured the feeling of ‘rejection’ (α = .91) using four statements (e.g., I feel rejected when I think about how I told statements (e.g., I feel inferior when I think about how I told (did not tell) about my vague suspicion).

Analysis of the data

PASW 23 (Predictive Analytics Software), previously called SPSS (Statistical Package for the Social Sciences) was used for the analyses. The Pearson correlation (two-tailed) was performed to determine the linear relationship between the two dependent variables. We used p < .05 as a measure of significance. This value helps to shed light on whether or not the results are due to chance (Fekjær, 2016). If the significance value is lower than .05, there is less than a 5% probability of the results being due to chance (Fekjær, 2016).

We also carried out one-way variance analyses (ANOVA) using PASW 23. This was done for the variables ‘seriousness’ and ‘acceptance’ in order to test the similarity between the two conditions (Field, 2013). We also carried out multivariate variance analyses (MANOVA) using PASW 23. We included multiple dependent measures for appraisals, feelings and motivations in separate analyses. This was done because Gausel and Leach (2011) show that the different theoretical variables are interlinked and are related to each other. MANOVA analyses are used to test the difference between groups in relation to several dependent variables at the same time, and we reported Wilks's lambda. Lambda is a measure of the percentage variation in the dependent variables that is not explained by the difference levels, for example self-appraisals, in the independent variables (Field, 2013).

Cohen's d was used as a measure of the effect size when the average and the standard deviation in the different conditions were compared. We carried out this effect measurement using an ‘effect size calculator’ that is available online (Becker, 1999). These results are shown in the table of average values. According to Cohen et al. (2003) d = .02 is regarded as a small effect size, d = 0.5 is regarded as a medium effect size, and d = .08 is regarded as a large effect size. This can mean that high d values show that a study has large average differences between the different conditions in the study.

Results

Experimental effects

Acceptance colleagues

An ANOVA demonstrated that the manipulation was not significant in relation to acceptance by colleagues F (1, 67) = .63, p = .43, partial ŋ2 = .009. ‘Tell’ (M = 6.07, SD = 1.01). ‘Not tell’ (M = 5.85, SD = 1.26). The pairwise comparison of acceptance by colleagues was not significantly higher p = .43, in the ‘tell’ condition (M = 6.07, SD = 1.01) than in ‘not tell’ (M = 5.85, SD = 1.26).

Acceptance family

The ANOVA results were not significant for acceptance by the family either F (1, 68) = .63, p = .27, partial ŋ2 = .02. ‘Tell’ (M = 5.05, SD = 1.6). ‘Not tell’ (M = 5.45, SD = 1.44). The pairwise comparison of acceptance by family was not significantly higher p = .27, in ‘not tell’ (M = 5.45, SD = 1.44) than in the ‘tell’ condition (M = 5.05, SD = 1.6). See also Tables 1 and 2 for the averages, standard deviations, Cohen's d and correlations for the different variables presented in the results below.

Table 1

Correlation and descriptive statistics.

	Variable	1	2	3	4	5	6	7	8
1	Seriousness	-
2	Appraisals Colleagues	.67*	-
3	Appraisals Others	.72*	.87*	-
4	Appraisals Self-image	.82*	.68*	.68*	-
6	Shame	.91*	.74*	.79*	.82*	.82*	-
7	Rejection	.66*	.73*	.78*	.69*	.66*	.78*	-
8	Inferiority	.80*	.68*	.76*	.76*	.79*	.86*	.84*	-
	Mean	3.6	2.8	3.1	2.3	4.3	2.9	2.4	1.5
	Standard deviation	2.3	1.6	1.7	1.5	1.8	2.0	1.4	1.5
	α	.98	.90	.94	.73	.58	.97	.91	.75

Note: N = 71. The response scale ranged from 1 (not at all) to 7 (strongly agree). *p < .05.

Table 2

Averages and standard deviations for appraisals (self-appraisals), seriousness, feelings and motivations.

	Tell		Not tell		Cohen's d
Variable	M	SD	M	SD	2v1
Appraisals Colleagues	1.88a	.75	3.85b	1.6	1.57
Appraisals Others	2.04a	.97	4.3b	1.5	1.78
Appraisals Self-image	1.3a	.48	3.5b	1.27	2.29
Seriousness	1.49a	.61	5.5b	1.34	3.85
Shame	1.3a	.53	4.7b	1.57	2.90
Rejection	1.59a	.68	3.21b	1.4	1.04
Inferiority	1.53a	.73	3.66b	1.34	1.97

Note: The average in each row that does not share the same letter is significantly different p < .05.

A MANOVA analysis showed a general effect of the vignette manipulation on the participants' appraisals relating to their concern about ‘social self-image colleagues’, ‘social self-image others’ and ‘self-image’, F (3, 67) = 38.04, p < .001, partial ŋ2 = .63. There was a significant univariate effect on appraisal colleagues, F (1, 69) = 45.13, p < .001, partial ŋ2 = .40. The participants in the ‘not tell’ condition expressed significantly higher average values (p < .001) for social self-image colleagues (M = 3.85, SD = 1.6) than the participants in the ‘tell’ condition (M = 1.88, SD = .75). There was a significant univariate effect on appraisal others, F (1,69) = 55.14, p < .001, partial ŋ2 = .44. Participants in the ‘not tell’ condition (M = 4.3, SD = 1.5) expressed significantly higher average values (p < .001) than in the ‘tell’ condition (M = 2.04, SD = .97) for appraisal others. There was a significant univariate effect on appraisal self-image, F (1, 69) = 99.75, p < .001, partial ŋ2 = .59. Participants in the ‘not tell’ condition (M = 3.5, SD = 1.27) expressed significantly higher average values (p < .001) than in the ‘tell’ condition (M = 2.04, SD = .48) for appraisal self-image.

An ANOVA analysis demonstrated that the manipulation had a significant univariate effect on seriousness F (1,69) = 276.1, p < .001, partial ŋ2 = .80. The pairwise comparison of seriousness was significantly higher p < .001 in the ‘not tell’ condition (M = 5.5, SD = 1.34) than in ‘tell’ (M = 1.49, SD = .61). See also Tables 1 and 2 for the averages, standard deviations, Cohen's d and correlations for the different variables presented in the results.

Feelings (shame, rejection and inferiority)

A MANOVA analysis showed a general effect of the vignette manipulation on the participants' feelings relating to their concern about 'felt shame', 'felt rejection' and 'felt inferiority', F (4, 65) = 38.04, p < .001, partial ŋ2 = .70. There was a significant univariate effect on felt shame, F (1, 68) = 154.17, p < .001, partial ŋ2 = .69. Participants in the ‘not tell’ condition (M = 4.7, SD = 1.57) expressed significantly higher average values (p < .001) than in the ‘tell’ condition (M = 1.3, SD = .53) for felt shame. There was a significant univariate effect on felt rejection, F (1, 68) = 37.7, p < .001, partial ŋ2 = .36. Participants in the ‘not tell’ condition (M = 3.21 SD = 1.4) expressed significantly higher average values on felt rejection (p < .001) than in the ‘tell’ condition (M = 1.59, SD = .68). There was a significant univariate effect on felt inferiority, F (1, 68) = 70.7, p < .001, partial ŋ2 = .51. Participants in the ‘not tell’ condition (M = 3.66, SD = 1.34) expressed significantly higher average values on felt inferiority (p < .001) than in the ‘tell’ condition (M = 1.53, SD = .73). See Tables 1 and 2 for the averages, standard deviations, Cohen's d and correlations for the different variables presented in the results.

Discussion

The purpose of the study was to explain the emotional discomfort experienced when the participants envisaged ‘telling’ or ‘not telling’ about vague suspicions of violence. In line with our hypothesis, we found support for the supposition that participants who ‘did not tell’ about vague suspicions of violence perceived this as being significantly more serious, and reported significantly higher intensity of emotional discomfort, than participants who ‘told’ about vague suspicions of violence. The participants in both conditions reported that acceptance by both the family and colleagues was very important. The results showed that participants in both conditions were more concerned with acceptance by colleagues than acceptance by the child's family, although the results were not significant. This indicates that acceptance is a highly valued need among the participants. This was not surprising since we know that the need for acceptance is a fundamental psychological need and that it is very important in terms of maintaining social bonds (Maslow, 1970). We assume that the results of our study could have a bearing on whether school or kindergarten staff choose to report vague suspicions that a child is being subjected to domestic violence (Maslow, 1970). Employees can choose not to report a vague suspicion in order to avoid breaking a social bond, although, if is later emerges that important information has been withheld, that can also damage the social bond. By telling about their vague suspicion, a school or kindergarten employee risks damaging a social bond, perhaps with a colleague who disagrees, or damaging a social bond with the family of the child in question (Scheff, 1994). Nevertheless, reporting "violence in grey areas" could be logically related to unclear definitions and protocol, and lack of training as well.

The results also show a significantly higher level for appraisals among the participants under the ‘not tell’ condition than the ‘tell’ condition. The participants in the ‘not tell’ condition show more concern about their self-image and social self-image than participants in the ‘tell’ condition. The actions of school staff are influenced by assessments they have made before the situation arises. Such assessments can influence whether an employee tells or does not tell about his or her vague suspicion because they are afraid of damaging a social bond, or of not being accepted by colleagues or the family of the child (Scheff, 1994; Maslow, 1970). It also depends on whether the employee risks something in the situation, for example if several values that the employee values highly are in conflict with each other (Lazarus, 1991).

Moreover, what employees envisage will happen when they report vague suspicions of violence will influence their choice of whether to tell or not tell. Negative reactions to telling about a vague suspicion can lead to an employee putting off reporting such suspicions. This is in line with Kenny (2001), who found that some teachers did not report suspicions because they were afraid of being mistaken. However, Webster et al. (2005) show that, with increased knowledge about the topic and an understanding that it is important to report such matters, more school employees will choose to tell about a vague suspicion. Moreover, this can be about an employee's values coming into conflict with each other (e.g. honesty and trust) and values that are valued more highly than other values (e.g. not hurting others) (Lazarus, 1991; Øverlien and Moen, 2016).

The results indicate that the participants believe that it is more serious not to report a vague suspicion of violence than to report it. This can be related to the fact that the participants are aware of the seriousness of a child living with domestic violence and that it can have negative consequences for the child both in the present and in future (Øverlien and Moen, 2016). Since we expect the majority of the respondents to be aware that the duty to report is a statutory duty, it is conceivable that this will influence how serious the respondents believe it is not to report such suspicions (The Child Welfare Act, 1993; The Convention on the Rights of the Child, 2003). For the majority of the participants in the study, the values that support reporting a vague suspicion were more important than negative reactions from the parents of the child in question or from colleagues (Lazarus, 1991).

Moreover, the results show a significantly higher level of self-critical feelings such as shame, inferiority and rejection among participants in the ‘not tell’ condition than in the ‘tell’ condition. Based on previous studies and our hypothesis, we expected to find more emotional discomfort among the participants as a result of not telling about their vague suspicions (Torp Løkkeberg, 2016). As mentioned earlier, the results show that the participants in the ‘not tell’ condition reported a higher level of self-critical emotions. This may be because the participants in the study know that it would have felt uncomfortable, i.e., they would have felt shame, and felt rejected and had feelings of inferiority in relation to colleagues, and also the family of the child, if they did not report their suspicion.

The aim of our study was to shed light on the emotional discomfort school staff and those training to become school or kindergarten staff can feel when they have a vague suspicion that a child is being subjected to domestic violence. In order to explain what emotional influence the different conditions can exert on someone who has a vague suspicion, we have looked at whether there is a difference between the group that was going to ‘tell’ and the group that was ‘not going to tell’ about their vague suspicion. The data in our study are consistent and show that there is a difference between the conditions. The data from the ‘not tell’ condition generally have a higher degree of intensity than the data from the ‘tell’ condition. This means that those who did not tell about their vague suspicion experienced greater discomfort than those who did tell about their suspicion.

Limitations of the study

This study has several limitations that we wish to discuss. The first limitation concerns the ‘not tell’ condition, where the participants were asked to envisage not telling about a vague suspicion of violence. For some participants, it can seem unnatural to envisage a situation in which, in real life, they would have acted in the opposite manner, and this could result in them disagreeing with the statements. Vignette studies based on questionnaires do not include interaction with and feedback from respondents, which are a necessary part of life (Hughes, 1998). Moreover, we wish to point out that it has been argued that vignettes are not directly comparable with real life (Faia, 1980). Vignettes nonetheless cover a broad range of relevant factors (e. g., social stimuli are standardised) and increase the external validity (Atzmüller and Steiner, 2010). The second limitation concerns the number of statements in the questionnaire. It is conceivable that reading and focusing on the different statements can be demanding for some participants. We believe that this has been taken into account by including a manipulation check in the questionnaire. Thirdly, when we measured experiences relating to shame, we used synonyms for shame to identify the feeling. We used a recognised, validated questionnaire, where, for example, we used ‘I feel ashamed’ and ‘I feel humiliated’ to measure the feeling of shame. It is important to emphasise that the vignettes do not reflect reality, and that shame is a very complicated feeling than many people may deny or not recognise, or not know what it feels like. It could threaten the validity if it is uncertain whether the respondent can relate to the feeling.

Summary and further follow-up studies

In this study, we have examined emotional discomfort among school staff when they have to tell about, compared with not telling about, vague suspicions that children are being subjected to domestic violence. In line with our hypothesis, we note that there are clear differences in how the respondents perceive themselves depending on which condition they are assigned to. Furthermore, in line with our assumption, we find that more emotional discomfort is associated with not telling about vague suspicions of violence than with telling about them. There is a clear need for educational institutions to raise this topic with current and future employees in the educational system. There also appears to be a need for more teaching about symptoms resulting from violence against children and adolescents in order to make it easier for school staff to detect violence and report their vague suspicions. In this way, school staff members can help prevent child abuse. The study indicates that school staff will find their situation less uncomfortable if they actually report vague suspicions that a pupil is a victim of domestic violence. This has great significant for safeguarding the school employees to choose to tell about vague suspicions such as inadequate clothing, signs of anxiety or lack of sleep, being verbally violent to peers or personnel, or reactions of fear towards sudden movements. If not, children can be stuck in the ugly situation of domestic violence.

Declarations

Author contribution statement

S. Torp Lokkeberg: Conceived and designed the experiments; Performed the experiments, Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

A. Balabanian, A.B. Øverby Aaseng, Aaseng, S. Økland: Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Competing interest statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.